Axiom Home Page Forums General Discussion The Water Cooler MP3 Quality test

A little disappointed, yet pleasantly suprised.

I ripped a store bought CD into my computer at 128 bps MP3 format. I then burned a CD of the album in MP3. The album I chose was from the band a perfect circle due to the large range of dynamics of the music and the high quality of the recording.

I placed the original in my first DVD playerand the MP3 version in my other DVD player. I started both discs simultaneously and then switched the sources back and forth. Well I was sure I was going to hear a difference and at first I could. The MP3 disc was quiter. I utilized my Onyko's intellivolume feature to match the loudness of both sources and then I couldn't hear any discernable difference. I was really hoping that my ears would be able to detect the loss of quality due to compression that I have been telling people exists when you go to MP3, but I just couldn't hear it. I listened on both a nice set of headphones and with my lovely axioms. They sounded the same. Which is good news, but I just thought I would be able to tell the difference.

Anybody done any similar tests with MP3 quality??

Not quite to the detail that you have, since I don't have the equipment. I have ripped some tracks at various encoding levels though and really haven't been able to hear the difference from the source once I got up above 128kbs or so. I guess that's a good thing, since it saves on disk space!

I rip everything at 320kbs and have A/B'd to the original disc with a ton of people and nobody to date can pick the disc from the MP3 or vice versa. Only problem with ripping at 320kbs is the required hard disc space. I'm currently running two 200gig hard drives and one is 85% full.

Neverhappy, how did you come to choose 320kbs? Did you have problems or notice degradation with lower bitrates?

That is what I ripped a bunch of stuff into my computer at as well but it seems that even at 128 you can't tell the difference.

I've noticed more sibilance and high-freq noise with some mp3's burned to cd when compared A/B to the original. Not all though. Also some seem to be, I don't know, drier. I rip at 320. Now with mp3 files on cd, I can hear quite a difference, like the atmosphere has been sucked out of the music.

I listen to nothing but mp3's. My Mazda's have come with kenwood mp3 players since 2001 as well so even in the car, it's all mp3.

Anything below 192 bit, I don't waste my time with. 128 is really poor quality, and I can easily tell the difference between it and a cd audio. It sounds tin'y, sibilance is more, less bass, almost muddy etc.

320 is great, but file size is a problem. So I try to get nothing but 192's or around 192. I think it's a good compromise. The difference between 192 to 320, at least to my ears is min.

Hmm... I find it strange you can't tell the difference. I can tell the difference even with 320 kbps MP3s when there's a lot going on in the music.

Higher frequencies are mainly where the flaws show up. Cymbals are especially good at revealing MP3 compression artifacts. I've never really heard cymbals sound correct. I can also readily tell with piano, but I am willing to guess that this is because I played the piano for 12 years.

If I may ask, how old are you? As you age, your ability to hear higher frequencies drops dramatically. I am guessing that people who can't tell the difference would be older than I am or have ears that have been abused by loud music.

Personally, I can't stand listening to MP3s when I'm actually paying attention to my music (lights off, door closed, eyes closed, and so on). The artifacts just jump out at me.

EDIT: newf and NeverHappy: that is what VBR with -q 11 in LAME is for. It gives you the same quality as CBR 320 but with nice space savings when a lower bitrate can be used without degrading the quality

I have always done everything everything with Exact Audio at 320kps and like I said, NOBODY has yet to sit in my house and pick out which is which.

I went 320kps only because I have always had big hard drives and anything below 320kps I seem to be able to pick out some things I don't like. 128kps for example sounds not bad but on some stuff a lot of bass that is normally there is gone or on some music it seems to add bass that was never there.

At the end of the day this like so many other things in audio comes down to personal taste and what the music is being played through. If you have a revealing system, chances are better you will be able to pick out the flaws associated with each respective bit rate.

.............and I'm 38 years old and just had my hearing checked. According to the Doc, it's as good as when I tested with him at 22. So far, so good.

I can hear the difference between 192 and 320. Just that I make cd's for the car, and I make the compromise at 192 so I can get more on a cd. So generally, all mp3's are ~192 or better.

I am 30

I found this on Wikipedia and thought it was interesting, especially the "Quality of MP3 Audio" and the suggestion to use "Lossless compression algorithms". It's a good and informative read, I really enjoyed it.

Here's the direct link to the relevant section: Wikipedia MP3

In reply to:

---

Hmm... I find it strange you can't tell the difference. I can tell the difference even with 320 kbps MP3s when there's a lot going on in the music.

---

I find it strange as well. I thought for sure that I would be able to tell the difference as I felt I had very discerning ears. Hence why I have invested so much time and money into my audio equipment. I am 31 years old and test just fine with my hearing. I am going to have to do more tests before I decide on what rate to go with but perhaps the burners, software, and media all play a part in this as well.

Or I just can't tell the difference.

Also for those of you that are certain that you can hear the difference have you done any a/b testing where you don't know which is being played? Cause I would have bet money that I could have guessed the different types before I actually did the test.

Alan if you are listening have you any experience with blind test of different mp3 formats?

This is all very interesting to me and thanks for all the comments guys.

for those who dont want to go all out with the 320kbps there is vbr, if utilized corretly. i use exact audio copy in secure mode with the alt lame extreme preset with yeilds bitrates ranging from 192-320, but it mostly stays in the 230-260 range, i cant difference between source and original with this bitrate, and it is quickly becoming the standard for release groups, rather than 192

OK. Just did another test with the same Perfect Circle album in both original and 128MP3. Playing at a level of 90-93db via direct mode (just M60's) and stereo mode (M60's and Hsu VTF3 MK2) I tried the test a couple different ways. One was to play a section of music on one disc, say the intro, a chorus, or a verse, then play the same section on the other disc. Then I tried the original test again of starting both discs simultaneously and switching back and forth.

Well here is what I have to say. . . .



Let it be known that this is an open invite. If anybody wants to come over my house and have a listen and show me that they can tell which is MP3 and which is the original as I switch them back and forth, I will reward them with much beer and much envy. For I cannot, and I cannot believe that I cannot.

More testing is to follow.

In reply to:

---

Let it be known that this is an open invite. If anybody wants to come over my house and have a listen and show me that they can tell which is MP3 and which is the original as I switch them back and forth, I will reward them with much beer and much envy. For I cannot, and I cannot believe that I cannot.

---

I have been doing this since the day's of Napster and to date and I'm not kidding, none of my beloved audiophile buddy's have been able to pick them apart when push comes to shove. I doubt you will have any takers that will prove you wrong.

In reply to:

---

I have been doing this since the day's of Napster and to date and I'm not kidding, none of my beloved audiophile buddy's have been able to pick them apart when push comes to shove. I doubt you will have any takers that will prove you wrong

---

We are talking about MP3s burned onto a music cd, not mp3s burned as files? As I said before, I can hear the difference on most MP3s, but only occasionally when they are burned as a music cd.

Either or. I have yet to meet someone who could pick it out whether I'm playing an MP3 at 320kps or the actual cd.

In reply to:

---

We are talking about MP3s burned onto a music cd, not mp3s burned as files? As I said before, I can hear the difference on most MP3s, but only occasionally when they are burned as a music cd.

---

I am ripping the original into the computer at 128 MP3 and then burning this album onto a blank cd-r media as an MP3 file but I don't think it would matter if I burned it as a Wav file because it has already been compressed and no amount of "uncompressing" should make it sound any better. 2nd law of thermodynamics or computerdynamics or something.

Hmmm... It's true that after ripping using lossy compression algorithms you don't recreate the lost info when burnig back onto a cd as a music/wav file. But last night, just trying to see if I'm fooling myself, I had my wife play "Us and Them" from DSOTM by Pink Floyd for me in 3 formats:

1) Original source, the remastered 20th Anniversary version.
2) MP3, ripped at 320kbps, burned as a file
3) MP3, ripped at 320kbps then burned as a music cd/wav.

And the one I picked out as different was the mp3 file. This wasn't especially scientific but I heard a difference immediately. On the other hand, I wrote down "Original" for the burned wav file and vice versa with the original source.
Again, the mp3 file was harsh on the highs and very dry overall. Could be my player? Who knows?

Very interesting. How much of the song were you playing at a time and did you have to change discs in the player, (ie how long of an interuption between one disc and another)?

I had to correct the volume level of the MP3 to match the other, was your MP3 stuff at the same db level as the original? (this is probably specific to my player setup situation but you never know, obviously people would pick out a volume level difference and could interpret this as quality. . .ie the bass and treble not hitting like they should)

Thanks for testing this out. Hope more can also so we can get some kind of consensus on the MP3's

CD's are in a changer. I listened to the whole song for all of them. My wife ran the changer. I used my trusty Rat Shack SPL meter to set the level at 90db for each disc. I waited 5 minutes between tests, which also allowed my wife to queue and pause the discs. Again, the results are probably dependent on many things. I used Grip to rip/encode, cdrdao to burn. It's possible my player isn't as good at reading mp3 files as others. I have noticed some recordings sound different between original source and ripped/burned as music discs. I haven't tried any of this using one of the lossless file formats. I have used Lame, Music Match, and WMP to rip, not really noticing any differences, then again I haven't done any 1:1 comparisons of the software as far as music quality. And I certainly don't claim to have golden ears that can hear the creamy-smooth richness of the rhodium-plated contacts of a $2500 power cable vs. the slightly sucked out sound of the $1500 one and the unlistenable harshness of the Rat Shack special. This test is certainly easy enough for all of us to try. I posit that the player is the most likely cause of real differences.

Hi all,

Thanks to sssSutherland for prompting me to post. I don't get down to the Water Cooler as often as I'd like.

Anyway, the blind tests I did with a number of other editors some years ago tested about ten different "lossy" algorithms, including MP3, running at various data rates.

Here are some of the conlusions I reached (and many of the other critical listeners):

At 320 kbps, the best of the compression algorithms are amazingly transparent. It took repeated A/B/X comparisons to isolate artifacts. DJ_Stunna is correct in stating that cymbals and, we found, other percussion instruments are really revealing of flaws. In particular, castinets were very revealing of artifacts. So, no flamenco at 320 kbps!

Another tip-off was increased sibilance on female vocals. We used a Suzanne Vega a cappella solo vocal (really annoying after many repetitions!) that became quite coarse and sibilant at slower data rates. If memory serves, it was also audible at 320 kbps.

One of the algorithms auditioned, developed for the Musicam digital radio system in Europe, was really excellent. I think it came out of that Fraunhoffer Institute in Germany. It was the one to be adopted for digital radio service in Canada. Sadly, the Canadians picked a part of the spectrum for the digital service that the Pentagon used for military purposes, so that aspect of digital radio in Canada never got beyond experimental broadcasts (which I heard; it was excellent).

Rock music wasn't as generally revealing of digital algorithm artifacts as classical, but then again in classical stuff you get a big variety of percussion instruments, the timbre of which may be slightly altered by a poor or too slow algorithm. Put another way, you don't get many castinets in rock and pop music!

Regards,

Alan -

These were MP3 files recorded as data, as opposed to MP3's burned as "music cd's?" Have you noticed a difference between the two? I am beginning to wonder if what I hear may be caused by my CDP possibly not processing MP3 files cleanly more than anything else.

BlueStater,

These were MP3 files recorded as data on a large hard drive and played back along with all the other lossy algorithms. We could do instantaneous A/B/X comparisons of each algorithm vs. uncompressed data files. I've never made comparisons of MP3s burned as music CDs.

Thanks Alan. It's that last difference that is noticeable on my system and makes me think there is some funny business with my players' ability to process MP3 files burned as data. As talked about earlier, an MP3 burned as a file and the same file burned as a "music" file should sound the same, being that the the bits are the same, so to speak.

In reply to:

---

Rock music wasn't as generally revealing of digital algorithm artifacts as classical,

---

Allright I'm going to have to test with some classical